Installing Custom Packages in Jupyter Notebook RRS feed

  • Question

  • I have an R script that fails with a strange error message in Azure ML Studio (no applicable method for 'roc' applied to an object of class "c('integer', 'numeric')") with code that works perfectly locally (Visual Studio).  Thus, I opened the same dataset in a Notebook to try to debug it.  In Jupyter, however, the versions of the packages in question installed with "install.packages" were quite different from those installed through a zip file in ML Studio.  

    1. How can I install custom versions of R packages in the Jupyter R textbook called from the Azure ML Datasets view?
    2. How can I find more debugging information on the error message from the Experiments view? I don't even know which package provides the message at this point.

    Tuesday, July 18, 2017 12:16 PM